MO-FG-204-06: A new algorithm for gold nano-particle concentration identification in dual energy CT

L. Chen, C. Shen, Kwok Po Ng, Tieyong Zeng, Y. Lou, X. Jia

Research output: Contribution to journalJournal article


Purpose: Gold nano-particle (GNP) has recently attracted a lot of attentions due to its potential as an imaging contrast agent and radiotherapy sensitiser. Imaging the GNP at its low contraction is a challenging problem. We propose a new algorithm to improve the identification of GNP based on dual energy CT (DECT). Methods: We consider three base materials: water, bone, and gold. Determining three density images from two images in DECT is an under-determined problem. We propose to solve this problem by exploring image domain sparsity via an optimization approach. The objective function contains four terms. A data-fidelity term ensures the fidelity between the identified material densities and the DECT images, while the other three terms enforces the sparsity in the gradient domain of the three images corresponding to the density of the base materials by using total variation (TV) regularization. A primal-dual algorithm is applied to solve the proposed optimization problem. We have performed simulation studies to test this model. Results: Our digital phantom in the tests contains water, bone regions and gold inserts of different sizes and densities. The gold inserts contain mixed material consisting of water with 1g/cm3 and gold at a certain density. At a low gold density of 0.0008 g/cm3, the insert is hardly visible in DECT images, especially for those with small sizes. Our algorithm is able to decompose the DECT into three density images. Those gold inserts at a low density can be clearly visualized in the density image. Conclusion: We have developed a new algorithm to decompose DECT images into three different material density images, in particular, to retrieve density of gold. Numerical studies showed promising results.
Original languageEnglish
Article number26852
JournalMedical Physics
Issue number6
Publication statusPublished - Jun 2015

User-Defined Keywords

  • Medical imaging
  • Gold
  • Nanoparticles
  • Computed tomography
  • Numerical modeling
  • Radiation therapy


Dive into the research topics of 'MO-FG-204-06: A new algorithm for gold nano-particle concentration identification in dual energy CT'. Together they form a unique fingerprint.

Cite this